Why small tools could survive the age of AI

Over the past few years, artificial intelligence has become capable of doing things that would have seemed quite impressive not so long ago. It can write code, explain an error, transform data, generate scripts, or even produce a (small) complete application from just a few instructions.

So it's easy to think that small tools will gradually lose their usefulness. After all, why use a converter or a generator when you can simply ask an AI to do the job?

And yet, I'm not convinced that's what will happen. I'll try to be as objective as possible (although I may be a little biased, I can't promise otherwise).

A matter of simplicity

Let's take a very ordinary example: a calculator. When I need to do a quick calculation, I very often use the calculator on my phone. And yet my phone is probably capable of doing much more than that. I could ask an AI to calculate the result. I could even write a small program to do it.

But I don't.

I simply open the calculator, type the numbers, and get the result.

That's probably one of the biggest strengths of small tools: they do one thing, and they do it immediately.

It's also why I regularly use online tools for very simple tasks. For example, I sometimes use a tool to encode or decode data, validate a format, or perform a small transformation.

And there's something a little strange about that. As a developer, my text editor is perfectly capable of doing some of these operations. I could probably achieve the same result with a terminal command or a few lines of code.

Yet I still end up searching for a small online tool. Not all the time, but sometimes I simply prefer that option.

AI sometimes adds a step

AI is incredibly versatile, but that versatility can also be a drawback. If I ask an AI to convert some data, I have to write a prompt, wait for a response, which is probably the part I find most annoying, and then verify that the result matches what I actually wanted.

With a specialized tool, the interface gives me exactly what I need straight away. This becomes even more obvious when the result has to be perfectly deterministic.

A converter, a validator, or a linter doesn't try to interpret my request. It applies precise rules. Given the same input and the same rules, I get the same output. Sometimes that's far more valuable than a generated response.

But is it still worth building these tools?

There's another paradox, though. Just because small tools remain useful doesn't necessarily mean it's worth building them yourself.

For a developer, creating a small tool can be very easy, especially with AI. But easy to build doesn't automatically mean worth building. If I only need to convert a few values once, writing my own script may take more time than using an existing tool.

That also reduces the incentive to create a personal version of something that already exists. Why maintain your own converter, manage its interface, handle bugs, and keep it updated when an online tool already does exactly what you need? Of course, if you want to be 100% confident about privacy, it can absolutely be worth it. Otherwise, it's a real discussion.

Small tools as shortcuts

So I think small tools could continue to have a place, but for a rather different reason than people might imagine.

They won't necessarily survive because they're more powerful than AI. They may simply be faster to use. A good small tool is almost a mental shortcut. I know what it does, I know where to click, and I know what to expect.

It doesn't need to understand a long description of my problem. That's probably why we still use calculators, stopwatches, unit converters, and validation tools even though our computers and phones are capable of doing infinitely more complex things.

In the end, the question may not be whether AI can replace these tools. It probably can in many situations. The real question is whether we actually want to replace a tool that takes five seconds to use with a conversation with an AI.

And in some cases, the answer will simply be no.